While doing T156174, I found this killgridjobs.sh. It checks if it is on bastion, but accepts only tools-bastion-01, which doesn't even exist anymore. As the script is only installed on bastions, I'd question the usefulness of this check.
Description
Details
Subject | Repo | Branch | Lines +/- | |
---|---|---|---|---|
toolforge: remove legacy killgridjobs.sh script | operations/puppet | production | +0 -44 |
Related Objects
- Mentioned Here
- T156174: Rewrite /usr/local/bin/crontab in python; fix bugs
Event Timeline
This has been superceded I believe by[[ https://github.com/wikimedia/operations-puppet/blob/35098e1e363e10696c2a19579fc141ff24480997/modules/toollabs/files/exec-manage | exec-manage ]] and can be removed.
https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Admin references neither exec-manage nor the scripts under modules/toollabs/files/gridscripts/ at the moment :-) (and the SOP in README in the latter does not seem to be documented on wikitech.wikimedia.org either). The procedure in README has courtesy mails to the owners of jobs to be deleted/rescheduled, I don't know if this is necessary/useful to keep.
I would suggest to wait for the next time a node needs to be drained, then document which scripts were actually run at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Admin#Draining_a_node_of_Jobs and afterwards remove everything unused.
Change 605345 had a related patch set uploaded (by BryanDavis; owner: Bryan Davis):
[operations/puppet@production] toolforge: remove legacy killgridjobs.sh script
Change 605345 merged by Bstorm:
[operations/puppet@production] toolforge: remove legacy killgridjobs.sh script
Mentioned in SAL (#wikimedia-cloud) [2020-06-15T21:28:27Z] <bstorm_> cleaned up killgridjobs.sh on the tools bastions T157792
Mentioned in SAL (#wikimedia-cloud) [2020-06-15T21:54:27Z] <bstorm_> removed killgridjobs.sh from toolsbeta bastion T157792